[R-sig-Geo] geoR help with kriging (newbie!)

[Roger Bivand]

On Thu, 9 Jun 2005, Doug Nychka wrote:

> Mike,
> Also check the package fields for Kriging it may produce output closer to
> your application. If the locations are the fields Krig function can take
> advantage of multiple fits ( your 300 cases) to different obsreeved
> values.
> Doug

Thanks to Doug and Paulo for replying so quickly; Mike's initial question
was posted on R-help about the fields package, and moved here after the
first reply there. I feel that Mike could try out some of the advice
offered to get a feel for which approaches will suit his needs, with
fields, geoR, and gstat, not to mention other packages, geostatistics is
now an interesting "field" to be in using R. 

We have tried to write our "Spatial" Task View (top left on your chosen
CRAN) to help beginners - suggestions for improving the Task View are very
welcome!

Roger

> 
> -----------------------------------------------------------------------------
> Doug Nychka,
> 
> Institute for Mathematics Applied to Geosciences
> National Center for Atmospheric Research
> Boulder, CO
> Email: nychka "AT" ucar "DOT" edu  Web: www.cgd.ucar.edu/~nychka
> Voice: 303-497-1711     FAX: 303-497-1333
> 
> 
> On Thu, 9 Jun 2005, Mike Smith wrote:
> 
> > Hi
> >
> > I have recently been experimenting with the use of kriging, primarily
> > through Goldensoftware's Surfer. Having had a look around at the
> > packages available for R I see the geoR (amongst several) supports
> > kriging and so have started experimenting with this.
> >
> > What I am interested in doing is batch processing ~300 datasets using
> > ordinary kriging with an exact interpolator (which I believe has no nugget).
> >
> > As I am not expert with either R or kriging (!!) I was wondering
> > what the best command line options might be? I have been working
> > through the geoR intro and have a few queries:
> >
> > -the command below appears to cover everything I want to do, however do
> > I need to fit a model to the variogram first and, if so, where is this
> > specified (using krige?)
> > kc <- krige.conv(s100, loc = pred.grid, krige = krige.control(obj.m = ml))
> >
> > -the input file is a data frame with three columns; x and y coordinates
> > and a z value to interpolate. I would ideally like to tell geoR (using
> > loc) to take the max extent of my area and, at a specified point spacing,
> > interpolate a new grid.
> >
> >
> > Many thanks
> >
> > mike
> >
> >
> > --
> > Dr Mike J Smith
> > Editor-in-Chief: Journal of Maps
> > http://www.journalofmaps.com
> >
> > School of Earth Sciences and Geography
> > Kingston University
> > Penrhyn Road
> > Kingston-upon-Thames
> > Surrey
> > KT1 2EE UK
> > tel +44 (0)208 547 7500
> > fax +44 (0)208 547 7497
> > OS X/Y:518098/168540
> > Lat/Long: 51.403603/-0.303575
> >
> > Personal Webpage:
> > http://www.kingston.ac.uk/esg/staff/smith.htm
> >
> > -----BEGIN PGP PUBLIC KEY BLOCK-----
> > Version: 2.6
> >
> > mQCPA0AyMQwBbgEEAMDTRNYB5Zi2SbmOtnBULHWIMXFaK+MVSTqGPi66N8s0ZQaJ
> > bMggYe5ySdltKLgT1FFmJN48IdSzqvzwPAMD5pQEuyLadnBfFvdfJapBcvHTd1pL
> > h5ikc/3zUXnK/JegioA4bCPypGnhf/1HFx70GRhpSaEOiul+vK1pQRliQJz1ABEB
> > AAG0KU1pa2UgU21pdGggPG1pY2hhZWwuc21pdGhAa2luZ3N0b24uYWMudWs+iQCV
> > AwUQQDIxDK1pQRliQJz1AQHn1QQAqA+203XQM7t0GEhIUh4PFEMysT+s2jMGUkNm
> > eihQnyK+1ljBcG9Vt8DtzOC5s3Pop1G9nbW9vl7OYzSb4E5uQ5L5wFUz0TJ81mu6
> > V8u/1CsERi+eXcNre+MON+q1GYoerUSd6LTh/kj+4J+MU05zHpSMXVJQsr2wcaxJ
> > vDM8DHY=
> > =oIrO
> > -----END PGP PUBLIC KEY BLOCK-----
> >
> > _______________________________________________
> > R-sig-Geo mailing list
> > R-sig-Geo at stat.math.ethz.ch
> > https://stat.ethz.ch/mailman/listinfo/r-sig-geo
> >
> 
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
> 

-- 
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Helleveien 30, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no